Microsoft's Webwright Agent Hits 60.1% on Odysseys Bench
Microsoft Research unveils Webwright, a terminal-native web agent framework that nearly doubles GPT-5.4's baseline score on the Odysseys benchmark, jumping from 33.5% to 60.1%.
Indonesia Pushes Real-Time Deepfake Defenses for Banks
Indonesian regulators are mobilizing stronger real-time deepfake detection across banking and fintech as synthetic identity fraud threatens the country's financial gates, pushing institutions toward continuous biometric and liveness verification.
GetReal Doubles Down on Continuous ID Verification
GetReal Security is expanding its continuous identity verification approach to combat increasingly sophisticated deepfake threats targeting enterprises, video calls, and authentication workflows.
DeepSeek Slashes V4 Pro AI Model Pricing by 75%
DeepSeek announces a permanent 75% discount on its new V4 Pro AI model, escalating the price war among foundation model providers and pressuring incumbents like OpenAI and Anthropic on API economics.
Reality Defender Expands Deepfake Defense Platform
Reality Defender is extending its deepfake detection platform into core cybersecurity workflows, hiring verification, and incident response — targeting three of the fastest-growing attack surfaces for synthetic media fraud.
AI Voice Cloning Resurrects Deceased Pilots' Voices
AI voice cloning technology is being used to recreate the voices of deceased pilots, raising new questions about synthetic audio, consent, and the boundaries of digital resurrection in aviation contexts.
AI Fabricated Quotes in Book, Author Defends Using It
An author discovered AI inserted fabricated 'synthetic quotes' into his published book, yet plans to continue using the technology. The incident highlights growing authenticity challenges in AI-assisted publishing.
AI Buildout Drives IT Capex to Record S&P 500 Share
IT capital expenditure has surged to a record share of S&P 500 spending as hyperscalers pour unprecedented sums into AI infrastructure, reshaping the economics of the entire tech sector and the compute foundation behind generative AI.
Microsoft Fara1.5 Agents Beat OpenAI, Gemini on Web Tasks
Microsoft's new Fara1.5 family of browser computer-use agents (4B/9B/27B) outperforms OpenAI Operator and Gemini 2.5 Computer Use on the Online-Mind2Web benchmark, marking a major step for open-weight web agents.
Edge LLMs Are Memory Bound: LiteRT Hits 30 Tok/s
Edge LLM inference is bottlenecked by memory bandwidth, not compute. Learn how LiteRT trades compute for bandwidth to achieve 30 tokens per second on resource-constrained devices through quantization and optimized memory access patterns.
SpaceX Bets Its Future on xAI as Grok Falters
As Grok struggles to keep pace with rivals, Musk is reportedly leveraging SpaceX's resources to bankroll xAI's bid to outcompete OpenAI, Google, and Anthropic — a high-stakes pivot with ripple effects across the AI ecosystem.
Trump Delays AI Security Executive Order to Avoid Stifling US Lea
President Trump has postponed signing an AI security executive order, citing concerns about hindering American leadership in artificial intelligence development.